← All posts
Ops   Jun 10, 2026 · 9 min read

Your top performer is your strategy's biggest single point of failure

Generated illustration for the post Your top performer is your strategy's biggest single point of failure

There is always one person. The Chief of Staff. The COO. The founder's right hand. The senior PM who's been there since year two. The one who knows, off the top of their head, which initiative belongs to which bet, which team owns which deliverable, why that thing in last quarter's plan got quietly dropped, who actually decided to slow the platform migration, and which dependency is about to slip on Tuesday.

They are indispensable. Everyone in the leadership meeting defers to them on questions of state. They are the reason the operating model holds together. They are also, structurally, the single largest risk in your strategy execution, and the company celebrates them rather than worrying about them.

Indispensability is a system failure, not a virtue

When the strategy-to-execution link lives in one person's head, three things are simultaneously true, and all of them are uncomfortable.

The company cannot scale that person's clarity beyond the rooms they are physically in. Whatever they understand about how the company actually runs is unavailable to anyone who isn't standing next to them when a question comes up. A new VP joins, a board member asks for context, a director needs to know whether her team's work is feeding the right bet — and the only path to the answer is a meeting with the one person.

A two-week holiday is a two-week visibility blackout. Decisions get deferred or made badly. Initiatives that needed adjustment don't get adjusted. Dependencies that should have been re-negotiated drift. When the person comes back, they spend the first week catching up on the damage that accumulated while they were away, which means the holiday cost the company three weeks of effective output, not two.

The day they leave, six months of institutional context leaves with them. Not because they're being uncooperative. Because most of what they know exists as muscle memory and Slack scrollback and Friday-afternoon conversations that never landed anywhere a successor could read. They write a handover doc. The doc covers maybe 20% of what they actually held. The rest is gone, and the new person spends two quarters rediscovering it by walking into avoidable mistakes.

We celebrate these people. We should be alarmed by the fact that they exist as a category.

What "the picture" actually is

When you ask the indispensable person to describe what they're holding, they often can't, because to them it feels like just knowing how things work. If you sit with them for an hour and pull on it, what comes out is a graph. Strategic bets at the top level. Objectives sitting under each bet. Initiatives sitting under each objective. Work items sitting under each initiative. Owners attached to every node. Dependencies running between nodes across team boundaries. Health states on every edge — what's on track, what's slipping, what's been quietly de-prioritised, what nobody has looked at in three weeks.

Most companies have this graph. It just doesn't exist as an artefact anyone can look at. It lives across one person's memory, three spreadsheets of varying staleness, two Notion docs that contradict each other, a Monday standup, a Friday review, and the dozen one-to-ones in between where the picture gets updated in fragments. The graph is real. It's just not durable. It depends entirely on the cognition of the person in the middle, and the moment that cognition is unavailable, the graph stops existing.

Making that graph a real artefact — durable, accessible to everyone with a legitimate reason to see it, updated as a side effect of the work rather than as a separate maintenance task — is the entire job of an execution layer. It is also the only way to fire the bus factor without firing the person.

The succession test

The cleanest test of whether your company has this problem is a simple thought experiment. Imagine your indispensable person quits on a Friday afternoon, walks out with their laptop, and never answers another message. On Monday morning, can the leadership team run the weekly review without them? Can the COO answer questions about the state of the top three bets from looking at your tools, not from memory? Could a competent new hire, parachuted into the role on day one, reconstruct the current state of the strategic bets, the health of each, who owns what, and what's being actively done about the at-risk ones — from the systems alone, without interviewing eight people?

If the honest answer is no, you don't have an operating system. You have a person, and you're calling them an operating system. That's a personnel risk masquerading as competence, and it gets more expensive every quarter you leave it unaddressed.

Why companies tolerate this

The indispensable person rarely complains about being indispensable. They're often proud of it. It's their value proposition, and asking them to systematise what they know feels, to them, like being asked to make themselves replaceable — which they correctly perceive as a threat to their leverage. So they resist, gently, and leadership defers to their resistance because they're the one person who can answer the question that just came up.

It's also rational from leadership's side in the short term. Building the artefact is a quarter of work that doesn't ship a customer feature, doesn't move a revenue number, and doesn't show up on the board deck. Continuing to rely on the person is free this week. The cost shows up only when something breaks — and by then, the cost is denominated in lost quarters, not in the engineering weeks it would have taken to fix it.

What changes when the graph is shared

When the strategy graph becomes a real artefact, the indispensable person doesn't become less valuable. They become differently valuable. They stop being the keeper of the state and start being the senior interpreter of it — which is the actual high-leverage role they should be in. They get to take a holiday. They get to coach the next layer. They get to focus on the questions where their judgement matters, instead of being a human cache for facts the system should be holding.

The company gets resilience. A new hire can come in and see the picture. A board member can be briefed in a meeting that doesn't require six prep hours from the chief of staff. A reorg doesn't reset the institutional memory because the memory wasn't in any one person to begin with.

The Vindaris view

The job of an execution layer is to make the person-in-the-middle redundant in the best possible way — by turning the strategy graph from private cognition into shared artefact. The person stays. Their judgement stays valuable. They stop being a single point of failure. If your strategic clarity depends on one human being in good health, on the right side of resignation, and physically in the room, you don't have an operating system yet. You have a hostage situation everyone is too polite to name.